Add http health check #210
Conversation
|
🤔 there was an issue with the vendored packages and when I did a tidy + vendored a massive set of changes were pulled in, taking a look to see what's up. |
|
Figured out the reason for the vendor issues. It was related to some indirect dependencies that is introduced by adding the We are running with 1.12 dependencies; I've opened #212 to look into updating our dependencies to 1.15. |
|
@timoreimann would mind taking another look? I've stripped down the health checking to not need new things vendored. |
|
I will backport this to the other branch PRs after this is thumbsed. |
| defer cancel() | ||
| _, _, err := c.account.Get(ctx) | ||
| if err != nil { | ||
| return fmt.Errorf("checking do health: %w", err) |
There was a problem hiding this comment.
nit: wrapping (%w) probably not needed.
(This will likely also fail on older versions built with Go <1.13.)
|
Conflicts need to be resolved still. |
|
Oh would be great if you could amend the change log real quick @nanzhong . |
We would like to be able to monitor the availability of the CSI driver. This PR adds a health check endpoint for this.
This is a similar change to digitalocean/digitalocean-cloud-controller-manager#293, but simpler since we have full control over how we configuring the driver.
Similar future consideration include: